The Log File Service commits the file operations every now and then and makes a checkpoint. If there are many concurrent transactions (Disk Activity) and the disk system cannot keep up, then the Log File Service will fall behind and the Recovery Log becomes full. The Log File Service will then block for new file operations while it commits the operations contained in the Recovery log to the disk. This blocking will reduce performance significantly.
If experiencing reduced performance when the computer is under stress, then one can check if it caused by a full Recovery Log:
- Configure Performance Logs and Alerts to monitor the activity of the Current Disk Queue Length counter for desired PhysicalDisk object.
- Observe if the Current Disk Queue Length is constantly high with intermittent drops to exactly 1
To see the current size of the Recovery Log for an NTFS partition:
To change the size (KBytes) of the Recovery Log for an NTFS partition:
CHKDSK <Drive:> /L
More Info MS KB101670
CHKDSK <Drive:> /L:65536 /F